The Neglected Significance of Glomerular Density as a 5-year Progression Indicator for IgA Nephropathy

Chin Med Sci J September 2017

Vol. 32, No. 3 P. 145-151

CHINESE MEDICAL SCIENCES JOURNAL ORIGINAL ARTICLE

The Neglected Significance of Glomerular Density as a 5-year Progression Indicator for IgA Nephropathy△ Zhenjie Chen, Hang Li, Jianfang Cai, Xin Zhang, Chao Li, Peimei Zou, Mingxi Li, Limeng Chen, Xuemei Li, Xuewang Li*, and Yubing Wen Department of Nephrology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing 100730, China

Key words: glomerular density; IgA nephropathy; renal progression Objective To investigate whether glomerular density (GD) could be an independent prognostic factor for patients of IgA nephropathy with estimated glomerular filtration rate (eGFR) of 30 to 60 ml/min per 1.73 m2, or for patients with time-average proteinuria < 0.5 g/d. Methods A total of 173 patients with biopsy-confirmed IgA nephropathy diagnosed from January 2000 to December 2010 were included. All of these patients were followed up for more than 5 years. The endpoint was a > 30% of decline in eGFR from baseline after 5-year follow-up. The optimal cut-off value of GD was calculated by ROC curve. Kaplan-Meier method and Cox regression analysis was used for survival analysis. Results A 30% of decline in eGFR occurred in 14.5% of all patients. The optimal diagnostic cut-off value of GD was 1.99/mm2 (AUC = 0.90, sensitivity = 84.0%, specificity = 81.8%) determined by ROC curve. The low GD group (GD < 1.99 per mm2) experienced a significant increase in renal endpoint for patients with eGFR of 30 to 60 ml/min per 1.73 m2 (six patients in lower GD group, while one patient in the other group). For patients with time-average proteinuria < 0.5 g/d, the lower GD group showed a higher eGFR decline from baseline (4.5±16.7 ml/min per 1.73 m2 vs. –8.1±21.4 ml/min per 1.73 m2, P = 0.038); two patients in this group reached the endpoint, while no patients in the higher GD group did. Conclusion GD could be an independent prognostic factor for patients of IgA nephropathy with eGFR at 30 to 60 ml/min per 1.73 m2 of body surface, particularly for those with time-averaged amount of urine protein less than 0.5 g per day.

Chin Med Sci J 2017; 32(3):145-151: DOI: 10.24920/J1001-9294.2017.021 Received for publication October 6, 2016. *Corresponding author Tel: 86-10-69158744, Fax: 86-10-69155088, E-mail: [email protected] △Supported by the Key Projects in the National Science and Technology Pillar Program During the Twelfth Five-year Plan Period (2011BAI10B03).

146

I

CHINESE MEDICAL SCIENCES JOURNAL

September 2017

GA nephropathy (IgAN) is the most common cause

following formula: MAP=DBP+1/3(SBP-DBP) (SBP: sys-

1

of primary glomerulonephritis all over the world.

tolic blood pressure, DBP: diastolic blood pressure). TA

Recent findings have shown that glomerular density

mean arterial pressure (MAP) and TA proteinuria were

(GD; nonsclerotic glomerular number per renal cor-

evaluated every 6 months during follow-up, and the aver-

tical area) can act as a quantitative parameter for kidney

age of all evaluations was used to represent the MAP and

disease. The mean (±SD) value of GD from 1395 living

proteinuria.6

kidney donors was 2.56 ± 0.9/mm2. In IgAN, GD might serve as an early indicator of long-term renal survival. A

Renal biopsy morphology

low GD was associated with a steeper decline in eGFR in

Serial sections of biopsy were stained with periodic acid-

IgAN patients with eGFR > 60 ml/min per 1.73 m.

How-

silver metheramine (PASM), and scanned into high-quality

ever, the cut-off value of GD in predicting renal survival is

digital slide images (Aperio CS2 system scanner, Leica Bi-

still unknown. In a primarily tubulointerstitial disease,

osystems Imaging, Inc. Nussloch, Germany). GD was de-

higher glomerulosclerosis percentage was associated with

fined as the following formula: GD = nonslerotic glomeruli

2-3

a lower GD and more serious renal damage. Since tubu-

number/profile area of cortex. Profile glomeruli area was

lointerstitial changes can be obvious in IgAN patients with

obtained directly by the image analysis software of Aperio

eGFR of 30 to 60 ml/min per 1.73 m , we want to deter-

scanner. Glomerular diameters were calculated using the

mine whether predicting with GD is still accurate in these

following formula:

patients. Although previous studies of IgAN showed that

glomerulus, d: diameter). Renal sections were also scored

patients with time-average (TA) proteinuria of <0.5 g/d

according to the Oxford Classification system.11-12 A par-

have relatively accurate predictions,5-7 the value of GD in

tially nonsclerotic glomerular tuft was counted as 0.5 com-

these patients was still difficult to predict. Therefore, we

plete one as in a previous study.2 PASM-stained sections

also aimed to find out whether the prediction exists in this

were magnified onto tablets to manually outline the cortex,

subgroup.

medulla, renal corpuscle volume, and nonsclerotic glomer-

4

2

= 2×

/

(GA: profile area of

ular tufts as previously reported.2

PATIENTS AND METHODS

Statistical analysis

Settings and participants

Baseline characteristics were divided into two groups

From January 2000 to December 2010, a total of 1607 pa-

based on the tertiles of GD. Numerical variables between

tients aged ≥18 years were confirmed by biopsy as having

groups were compared using two tailed independent sam-

primary IgAN at the Peking Union Medical College Hospital,

ple t tests, Mann-Whitney U tests, and Wilcoxon signed

Beijing, China. Many of these patients were excluded, in-

ranks tests, as appropriate. Categorical variables were

cluding 6 patients with rapidly progressive glomerulone-

presented as numbers (percentage) and compared using

phritis, 1012 patients with follow-up less than 5 years, 24

Chi-square tests.

patients who had fewer than 8 glomeruli in renal tissue

Here, Cox regression was used to estimate the hazard

specimens, and 13 patients with baseline eGFR ≤ 30

ratio (HR) for the renal endpoint. A Cox proportional haz-

ml/min per 1.73 m2. The left total of 552 patients were

ard regression model was used to assess unadjusted and

included in this study. Then 180 patients were randomly

adjusted associations between tertiles of GD value and

selected for further pathological analysis, among which, 7

renal endpoints. C-statistic was used to compare the ef-

patients who had undergone only irregular follow-up were

ficiency of risk prediction of the model with GD only to

excluded, leaving 173 patients for final analysis. All pa-

those of the models with clinical parameters, or with both

tients provided written informed consent for the renal bi-

clinical and pathological parameters. C-statistic is a

opsy. The study protocol was approved by the ethics com-

measure of the discriminative power of the logistic equa-

mittee of the Peking Union Medical College Hospital.

tion, which is calculated by running a ROC curve with the obtained predictive probabilities as the test variable. The

Measurements

c-statistic value is the area under the curve. Models are

The eGFR was calculated using the Chronic Kidney Disease

typically considered reasonable when the C-statistic ex-

Epidemiology Collaboration (CKD-EPI) formula.8 The renal

ceeds 0.7, and strong when it exceeds 0.8. ROC analysis

endpoint was defined as a >30% of reduction in the eGFR

was used to evaluate the diagnostic accuracy of nonscle-

from the baseline after 5-year follow-up.9-10 Mean arterial

rotic GD for renal endpoints. Univariate and multiple linear

pressure (MAP) was calculated approximately using the

regression analysis were performed to assess the relation

Vol. 32, No.3

CHINESE MEDICAL SCIENCES JOURNAL

147

between eGFR decline and continuous variables. Statisti-

terms of age, hypertension, eGFR, relative amount of

cal analyses were run using IBM SPSS Statistics software

global glomerulosclerosis, and maximal glomerular diame-

version 21 (IBM Corp., Armonk, NY, U.S.). A two-tailed P

ter (all P<0.05, Table 1). At the end, 25 (14.5%) patients

value < 0.05 was considered to be statistically significant.

had a 30% decline in eGFR during follow-up. Significance of GD in predicting renal end point among

RESULTS

the participants

Baseline characteristics of the study population

Four models were constructed to estimate HR for a 30 percent

As shown in Table 1, 74 patients, who had biopsy, were

decline in eGFR among the patients with GD equal to or less

male, with a median urinary protein of 2.0 (Q1, Q3, 0.9,

than 2.18/mm2 and those with GD greater than 2.18/mm2.

4.0) g/d and eGFR of 80.0 ± 25.4 ml/min per 1.73 m , by

The unadjusted HR was 12.41 (4.26, 36.18). When adjusted

the time of biopsy. By comparing variables of the group

for age and sex in model 1, it was 14.05 (4.79, 41.21). The

containing the first tertile with those of the group containing

HR was 13.14 (4.08, 42.39), 10.80 (3.03, 8.45), and

the last two tertiles, there were significant differences in

7.21(1.82, 28.52) in following models, as shown in Table 2.

2

Table 1. Comparisons of baseline characteristics between the group containing the first tertile and the group containing the last two tertiles of glomerular density at the time of biopsy in patients with IgA nephropathy Characteristics n Age§ (yrs)

All patients

Glomerular density ≤2.18/mm2

>2.18/mm2

173

58

115

36.2 ± 9.9

36.6 ± 9.3

36.0±10.3

t value

－0.25

P value

0.001

Male [n (%)]

74 (42.8)

25 (43.1)

49 (42.6)

0.54

Diabetes [n (%)]

15 (8.7)

3 (5.2)

12 (10.4)

0.19

Hypertension [n (%)]

71 (41.0)

31 (53.4)

40 (34.8)

0.014

Body mass index§ (kg/m2)

24.5 ± 3.9

24.3 ± 3.6

24.5 ± 4.1

－0.12

0.75

2.17

0.09

Mean arterial pressure§ (mm Hg)

96.0 ± 13.3

98.4 ± 14.2

94.8 ± 12.6

Proteinuria [g/d, median (Q1, Q3)]

2.0 (0.9, 4.0)

2.2 (1.0, 3.8)

2.0 (0.9, 4.0)

eGFR (ml/min per 1.73 m )

80.0 ± 25.4

67.9 ± 22.0

86.0 ± 24.9

－4.73

<0.001

Serum creatinine (μmol/L)

100.3 ± 39.8

113.0 ± 39.0

89.8 ± 27.8

3.99

<0.001

Uric acid (μmol/L)

361.0 ± 109.2

405.8 ± 114.2

338.3 ± 99.6

5.08

<0.001

Total cholesterol [mmol/L, median (Q1, Q3)]

5.6 (4.7, 6.7)

5.6 (4.7, 6.6)

5.5 (4.7, 6.8)

Triglycerides (mmol/L)

2.2 ± 1.2

2.1 ± 1.2

2.2 ± 1.2

－0.33

0.48

2.7 ± 1.1

1.5 ± 0.4

3.3 ± 0.9

－8.21

<0.001

§

2

§

§

§

0.94

0.94

Pathological variables Glomerular density§ (per mm2) Global glomerulosclerosis§ (%)

26.1 ± 18.4

40.8 ± 18.8

18.6 ± 13.0

8.19

<0.001

Maximal glomerular diameter§ (μm)

216.3 ± 25.8

224.1 ± 22.6

212.4 ± 26.4

2.57

0.004

M1 [n (%)]

102 (59.0)

35 (60.3)

67 (58.3)

0.46

E1 [n (%)]

52 (30.1)

15 (25.9)

37 (32.2)

0.25

S1 [n (%)]

163 (94.2)

56 (96.6)

107 (93.0)

0.29

T1 [n (%)]

58 (33.5)

21 (36.2)

37 (32.2)

0.36

T2 [n (%)]

17 (9.8)

14 (24.1)

3 (2.6)

<0.001

§: Plus-minus values are means±SD. eGFR: estimated glomerular filtration rate; M1: mesangial hypercellularity score > 0.5; E1: presence of endocapillary hypercellularity; S1: presence of segmental glomerulosclerosis; T1: tubular atrophy/interstitial fibrosis is 26%-50% of cortical area; T2: tubular atrophy/interstitial fibrosis is >50% of cortical area.

148

CHINESE MEDICAL SCIENCES JOURNAL

September 2017

Table 2. Results of Cox regression analysis for the risk of a 30% of decline in eGFR by using glomerular density and other variables in the patients with IgA nephropathy ([hazard ratio (95% CI)] Characteristics

Unadjusted

Model 1

Model 2

Model 3

Model 4

Glomerular density (per mm ) 2

>2.18/mm2 ≤2.18/mm2

1.0 (Ref)

1.0 (Ref)

1.0 (Ref)

1.0 (Ref)

12.41 (4.26, 36.18) 14.05 (4.79, 41.21) 13.14 (4.08, 42.39) 10.80 (3.03, 8.45)

1.0 (Ref) 7.21 (1.82, 28.52)

Age

0.94 (0.90, 0.99)

0.92 (0.87, 0.98)

0.93 (0.88, 1.00)

0.94 (0.89, 1.00)

Gender

1.49 (0.66, 3.34)

2.18 (0.85, 5.57)

2.11 (0.80, 5.58)

2.14 (0.81, 5.66)

Hypertension

0.80 (0.33, 1.80)

0.85 (0.32, 2.31)

0.90 (0.34, 2.38)

Proteinuria

1.12 (0.95 ,1.32)

1.09 (0.89, 1.34)

1.13 (0.91, 1.40)

eGFR

0.98 (0.96, 1.00)

0.99 (0.97, 1.01)

0.99 (0.97, 1.02)

Corticosteroids

4.38 (1.23, 15.53)

4.33 (1.07, 17.46)

6.16 (0.40, 27.11)

RAS blockers

0.65 (0.08, 5.53)

0.64 (0.07, 6.21)

0.84 (0.08, 8.76)

Immunosuppressive agents

2.27 (0.79, 6.55)

2.18 (0.70, 6.74)

2.08 (0.66, 6.57)

M1

1.07 (0.42, 2.73)

1.13 (0.44, 2.94)

E1

1.03 (0.40, 2.69)

1.19 (0.45, 3.20)

S1

0.87 (0.08, 9.98)

0.80 (0.07, 9.75)

T1-2

1.02 (0.98, 1.06)

1.01 (0.97, 1.05)

1.00 (0.98, 1.03)

1.01 (0.99, 1.03)

Maximal glomerular diameter Global glomerulosclerosis percent

1.02 (1.00, 1.06)

RAS blockers: renin-angiotensin system blockers. Model 1: adjusted for age and sex. Model 2: adjusted for the covariates in the Model 1 plus hypertension, proteinuria, eGFR, and use of corticosteroids, RAS blockers, and immunosuppressive agents (cyclophosphamide, mycophenolate mofetil, azathioprine, or cyclosporine A). Model 3: adjusted for the covariates in the model 2 plus M1, E1, S1, T1-2 and maximal glomerular diameter. Model 4: adjusted for the covariates in the Model 3 plus relative amount of global glomerulosclerosis. Ref indicating as reference.

Three models were used to compare the prediction efficiency in a sensitivity analysis by C-statistics, including one model with only GD, one with only clinical parameters, and one with both clinical and pathological parameters (Fig. 1). The results indicated that the model including both clinical and pathological variables had the highest C-statistics value (C-statistic = 0.99; 95%CI, 0.98-1.00). The ROC analysis showed the optimal cut-off value of GD to be 1.99/mm2 (area under the curve = 0.91, sensitivity = 85.2%, specificity = 81.2%). Here, 44 patients with eGFR of 30 to 60 ml/min per 1.73 m2 were analyzed for the prediction of GD. They were divided into two groups according to the cut-off value of GD (1.99/mm2). A comparison of baseline characteristics between the lower and the higher GD group is given in Table 3. Age in the lower GD group (GD < 1.99/mm2) showed younger than that in the higher GD group (t = －2.66, P = 0.01). No significant differences were detected in other parameters. Patients in the lower GD group experienced a significant increase in renal endpoint (six pa-

Figure 1. Receiver-operating characteristic curve for C-statistic in four different Cox models predicting the risk for eGFR decline in the patients with IgA nephropathy. The model with only clinical parameters included eGFR, mean arterial pressure over time, and proteinuria over time. The model with both clinical and pathological parameters included these clinical parameters, relative amount of segmental glomerul-

tients in lower GD group, while one patient in the other

osclerosis, tubular atrophy/interstitial fibrosis ≥26%,

group), as indicated by Kaplan–Meier analysis (Fig. 2).

and glomerular density.

Vol. 32, No.3

CHINESE MEDICAL SCIENCES JOURNAL

149

Table 3. Comparison of baseline characteristics of patients (n=22) with eGFR of 30 to 60 ml/min per 1.73 m2 at biopsy between the lower and the higher glomerular density group Glomerular density Characteristics

≤1.99/mm2

>1.99/mm2

Age§ (yrs)

33.8±9.2

41.7±10.4

Male [n (%)]

10 (45.5)

10 (45.5)

Body mass index§ (kg/m2)

23.3±3.6

25.4±4.4

Mean arterial pressure§ (mm Hg)

99.6±14.4

Proteinuria (g/d)

t value

P value

−2.66

0.01 1.00

−1.74

0.09

98.2±10.9

0.36

0.72

3.9±2.9

2.7±2.5

1.51

0.14

eGFR (ml/min per 1.73 m )

46.8±7.8

51.2±9.0

−1.72

0.09

Uric acid§ (μmol/L)

460.5±84.4

408.2±91.2

1.98

0.06

Total cholesterol§ (mmol/L)

6.3±1.6

5.7±1.5

1.37

0.18

Triglycerides (mmol/L)

2.3±1.3

2.6±1.2

−0.72

0.47

Maximal glomerular diameter§ (μm)

224.8±26.9

211.8±19.8

1.61

0.08

M1 [n (%)]

14 (63.6)

9 (40.9)

0.13

E1 [n (%)]

8 (36.4)

7 (31.8)

0.75

S1 [n (%)]

22 (100.0)

21 (95.5)

0.31

T1 [n (%)]

11 (50.0)

10 (45.5)

0.76

T2 [n (%)]

6 (27.3)

5 (22.7)

0.73

§

§

2

§

Pathological variables

§: Plus-minus values are means±SD.

per 1.73 m2, P = 0.038). Two patients experienced a 30% of decline in eGFR. Their GDs were all below 1.13/mm2. In univariate linear regression analysis with the eGFR decline (%/5 yrs) as the dependent variable, a significant association was noted between the eGFR decline and eGFR level at the time of biopsy (R=0.412, P<0.001). When adjusted by multiple linear regression, both eGFR level at the time of biopsy (standardized β=0.611, P<0.001) and GD (standardized β=–0.297, P=0.004) showed significant associations with eGFR decline.

DISCUSSION Figure 2. Comparison of cumulative renal survival rate between

Our results showed GD to be correlated with eGFR de-

the group of patients with glomerular density

cline. The cut-off value of GD in predicting the incidence of

(GD) >1.99/mm2 and the one ≤1.99/mm2.

a 30% of decline in eGFR during 5-year follow-up was 1.99/mm2. GD has a similar prediction efficiency compared

Prediction of glomerular density in patients with TA

with the combination of TA urinary protein, TA mean arte-

urinary protein of < 0.5 g/d

rial pressure and eGFR by C-statistic means. For patients

In our study, 74 patients had TA urinary protein of < 0.5

with eGFR of 30 to 60 ml/min per 1.73 m2, the lower GD

g/d, among whom there were 59 patients with GD levels

group experienced a significant increase in the incidence

higher than 1.99/mm . The lower GD group had a signifi-

of renal end point. Patients with a higher GD whose TA

cant decrease in eGFR decline (4.5±16.7 ml/min per 1.73

proteinuria level is < 0.5 g/d, also had a better prognosis.

m2) compared to the higher GD group (-8.1±21.4 ml/min

Glomerular sclerosis is a common cause of nephron

2

150

CHINESE MEDICAL SCIENCES JOURNAL

September 2017

loss. A positive correlation has been observed between low

In summary, GD had a good prediction for renal sur-

GD and the relative amount of glomerular sclerosis (r=

vival when defined as a 30% of decline in eGFR during a

–0.65, P<0.001). However, glomerular sclerosis showed

5-year follow-up; the higher GD group experienced a bet-

no prediction on renal survival by multivariate Cox analysis

ter prognosis. Lower GD was still predictive even for pa-

(data not shown). When adjusted by models including glo-

tients with well-controlled TA urinary protein and eGFR of

merular sclerosis, the group with a GD < 1.99/mm2 still

30 to 60 ml/min per 1.73 m2. The prediction model with

showed significantly higher HR than the group with a GD >

only GD was as sensitive as the model that used TA urinary

1.99/mm (Table 2). This can be explained by the fact that

protein, TA mean arterial pressure, and eGFR.

2

low GD and glomerulosclerosis could be separate processes, even though glomerular sclerosis plays an im-

Conflict of Interest Statement

portant part in leading to low glomerular density.13 Lower

The authors have no conflict of interest to disclose.

GD can also be caused by low birth weight,14-15 hypertensive nephrosclerosis,16 higher body surface, and older

ACKNOWLEDGMENTS

age.17-18

We thank LetPub (www.letpub.com) for its linguistic assis-

The lower GD group showed a significant increase in

tance during the preparation of this manuscript.

the incidence of renal endpoint in patients with eGFR less than 60 ml/min per 1.73 m2. Besides, GD can serve as a

REFERENCES

response index to corticosteroids therapy in adult patients with kidney disease.19 In this way, the treatment

1.

1.73 m has been controversial. KDIGO clinical practice 2

Wyatt RJ, Julian BA. IgA nephropathy. N Engl J Med 2013; 368(25):2402-14. doi: 10.1056/NEJMra1206793.

of corticosteroids in IgAN with eGFR of <50 ml/min per 2.

Elsherbiny HE, Alexander MP, Kremers WK, Park WD,

provided no guidance for the use of cortico-

Poggio ED, Prieto M, et al. Nephron hypertrophy and glo-

steroids in patients of IgAN with eGFR of <50 ml/min per

merulosclerosis and their association with kidney func-

guideline

20

1.73 m , while controversial results were obtained by

tion and risk factors among living kidney donors. Clin J

VALIGA study21 and STOP-IgAN study.22 Further prospec-

Am Soc Nephrol 2014; 9(11):1892-902. doi: 10.2215/

tive studies taking GD into account are needed to evalu-

CJN.02560314.

2

ate the benefit of corticosteroids in patients with a eGFR

3.

Tsuboi N, Kawamura T, Koike K, Okonogi H, Hirano K, Hamaguchi A, et al. Glomerular density in renal biopsy

of <50 ml/min per 1.73 m2. As for IgAN, patients with TA urinary protein of >1.0 g/d

specimens predicts the long-term prognosis of IgA

were associated with a 46.5-fold higher risk than those

nephropathy. Clin J Am Soc Nephrol 2010; 5(1):39-44.

with TA urinary protein of <0.5 g/d.5 The VALIGA study

doi: 10.2215/CJN.04680709.

also suggested that TA urinary protein of <0.5 g/d is cor-

4.

Hewitson TD, Darby IA, Bisucci T, Jones CL, Becker GJ.

related with a significantly better outcome than that of 0.5-

Evolution of tubulointerstitial fibrosis in experimental re-

1.0 g/d. The current study showed that GD still plays an

nal infection and scarring. J Am Soc Nephrol 1998;

6

9(4):632-42.

important role in the disease progression for patients with TA urinary protein of <0.5 g/d. Two patients experienced a

5.

Le W, Liang S, Hu Y, Deng K, Bao H, Zeng C, et al. Long-

decrease in eGFR of more than 30% from the baseline.

term renal survival and related risk factors in patients

Their GD were all less than 1.3/mm . The value of GD

with IgA nephropathy: results from a cohort of 1155

should be given particular attention for patients with TA

cases in a Chinese adult population. Nephrol Dial Trans-

2

plant 2012; 27(4):1479-85. doi: 10.1093/ndt/gfr527.

urinary protein of <0.5 g/d. One limitation of this study is its observational nature,

6.

Coppo R, Troyanov S, Bellur S, Cattran D, Cook HT, Fee-

so the predictive efficiency of renal survival is less powerful.

hally J, et al. Validation of the Oxford classification of IgA

Second, the case numbers of subgroups were relatively

nephropathy cohorts with different presentations and

small, and future studies should use larger samples. The

treatments. Kidney Int 2014; 86(4):828-36. doi: 10.

5-year study duration was the third limitation. It was

1038/ki.2014.63.

shorter than a previous study in which patients were fol-

7.

Sarcina C, Tinelli C, Ferrario F, Visciano B, Pani A, De Sil-

lowed up for 10 years.3 Patients whose follow-up periods

vestri A, et al. Corticosteroid treatment influences TA-

lasted less than 5 years were excluded. This could cause

proteinuria and renal survival in IgA nephropathy. PLoS

some selection bias with exclusion of patients who showed

One 2016; 11(7):e0158584. doi: 10.1371/journal.pone.

an eGFR decline of 30%.

0158584.

Vol. 32, No.3 8.

9.

CHINESE MEDICAL SCIENCES JOURNAL

Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF

Okabayashi Y, et al. Glomerular density and volume in

3rd, Feldman HI, et al. A new equation to estimate glo-

renal biopsy specimens of children with proteinuria rela-

merular filtration rate. Ann Intern Med 2009; 150(9):

tive to preterm birth and gestational age. Clin J Am Soc

604-12. doi: 10.7326/0003-4819-150-9-200905050-00006.

Nephrol 2017; 12(4):585-90. doi: 10.2215/CJN.05650516.

Coresh J, Turin TC, Matsushita K, Sang Y, Ballew SH, Ap-

16.

Shimizu A, et al. Glomerular density in biopsy-proven hy-

rate and subsequent risk of end-stage renal disease and

pertensive nephrosclerosis. Am J Hypertens 2015; 28(9):

jama.2014.6634.

1164-71. doi: 10.1093/ajh/hpu267. 17.

relation to age, kidney weight, and body surface in nor-

Lewis E, et al. GFR decline as an end point for clinical

mal

trials in CKD: a scientific workshop sponsored by the Na-

10.1002/ar.1092320205. 18.

1992;

232(2):194-201.

doi:

Tsuboi N, Utsunomiya Y, Kanzaki G, Koike K, Ikegami M, ulomegaly in obesity-related glomerulopathy. Clin J Am

Working Group of the International IgA Nephropathy Net-

Soc Nephrol 2012; 7(5):735-41. doi: 10.2215/CJN.07270711. 19.

Koike K, Tsuboi N, Utsunomiya Y, Kawamura T, Hosoya T.

R, Cook HT, Feehally J, Roberts IS, et al. The Oxford clas-

Glomerular density-associated changes in clinicopatho-

sification of IgA nephropathy: rationale, clinicopatholog-

logical features of minimal change nephrotic syndrome in

ical correlations, and classification. Kidney Int 2009;

adults. Am J Nephrol 2011; 34(6):542-8. doi: 10.1159/

Trimarchi H, Barratt J, Cattran DC, Cook HT4, Coppo R,

000334360. 20.

Radhakrishnan J, Cattran DC. The KDIGO practice guide-

Haas M, et al. Oxford classification of IgA nephropathy

line on glomerulonephritis: reading between the (guide)

2016: an update from the IgA nephropathy classification

lines—application to the individual patient. Kidney Int

working group. Kidney Int 2017; 91(5):1014-21. doi:

2012; 82(8):840-56. doi: 10.1038/ki.2012.280. 21.

Tesar V1, Troyanov S, Bellur S, Verhave JC, Cook HT, Fee-

Hoy WE, Bertram JF, Hughson MD. Nephron hypertrophy

hally J, et al. Corticosteroids in IgA nephropathy: a ret-

and glomerulosclerosis in normal donor kidneys. Clin J

rospective analysis from the VALIGA study. J Am Soc

Am Soc Nephrol 2014; 9(11):1832-4. doi: 10.2215/CJN.

Nephrol 2015; 26(9):2248-58. doi: 10.1681/ASN.20140

08680814.

15.

Rec

Kawamura T, et al. Low glomerular density with glomer-

10.1016/j.kint.2017.02.003.

14.

Anat

10.1053/j.ajkd.2014.07.030.

76(5):534-45. doi: 10.1038/ki.2009.243.

13.

man.

ministration. Am J Kidney Dis 2014; 64(6):821-35. doi:

work and the Renal Pathology Society, Cattran DC, Coppo

12.

Nyengaard J, Bendtsen T. Glomerular number and size in

Levey AS, Inker LA, Matsushita K, Greene T, Willis K,

tional Kidney Foundation and the US Food and Drug Ad-

11.

Haruhara K, Tsuboi N, Kanzaki G, Koike K, Suyama M,

pel LJ, et al. Decline in estimated glomerular filtration mortality. JAMA 2014; 311(24):2518-31. doi: 10.1001/ 10.

151

Luyckx VA, Brenner BM. Birth weight, malnutrition and

70697. 22.

Rauen T, Eitner F, Fitzner C, Sommerer C, Zeier M, Otte

kidney-associated outcomes—a global concern. Nat Rev

B, et al. Intensive supportive care plus immunosuppres-

Nephrol 2015; 11(3):135-49. doi: 10.1038/nrneph.2014.251.

sion in IgA nephropathy. N Engl J Med 2015; 373(23):

Koike K, Ikezumi Y, Tsuboi N, Kanzaki G, Haruhara K,

2225-36. doi: 10.1056/NEJMoa1415463.